1. Fit the saturated regression of the number of projects per 1,000 on the variables defined in the next 33 columns of the data table (Pct Added Housing Sq Ft through Income [per cap] 2000).
(a) Why is this model difficult to interpret?
(b) How accurately will this model predict new locations?
2. Apply forward stepwise regression to build a more parsimonious model from this collec- tion of potential explanatory variables. Be sure to avoid over-fitting. Interpret, if you can, the model obtained by stepwise. Is collinearity a problem?